Reimagining research papers as interactive and reliable AI agents.

Paper2Agent converts research papers into interactive AI agents, enabling complex scientific queries and collaboration with over 80% accuracy in reproducing original results.

• Why it matters: Current research dissemination relies on static documents that hinder reuse and understanding, creating barriers for scientists to efficiently build upon existing work. Transforming papers into active AI agents can accelerate discovery and collaboration across scientific fields.
• What they did: The framework analyzes papers and codebases using multiple AI agents to build a model context protocol (MCP), which is then tested and refined. It connects MCPs to chat agents like Claude Code to interpret data, run workflows, and answer complex questions.
• The result: This approach successfully reproduces original research results and enables novel user queries, fostering a new ecosystem of interactive, collaborative AI co-scientists that enhance knowledge sharing and discovery.